INTERNATIONAL JOURNAL OF ONCOLOGY 43: 457-468, 2013

The Wnt signaling pathway and mitotic regulators in the initiation and evolution of mantle cell lymphoma: expression analysis

YOSHIZO KIMURA1, FUMIKO ARAKAWA1, JUNICHI KIYASU1, HIROAKI MIYOSHI1, MAKI YOSHIDA1, AYAKO ICHIKAWA1, DAISUKE NIINO1, YASUO SUGITA1, TAKASHI OKAMURA2, ATUSHI DOI3, KAORI YASUDA3, KOSUKE TASHIRO4, SATORU KUHARA4 and KOICHI OHSHIMA1

Departments of 1Pathology and 2Hematology, School of Medicine, Kurume University, Kurume, Fukuoka; 3Cell Innovator Inc., Venture Business Laboratory of Kyushu University; 4Department of Bioscience and Biotechnology, Faculty of Agriculture, Kyushu University, Fukuoka, Japan

Received April 10, 2013; Accepted May 20, 2013

DOI: 10.3892/ijo.2013.1982

Abstract. For an accurate understanding of mantle cell β- with high levels of cytoplasmic Wnt3 staining. For lymphoma (MCL), molecular behavior could be staged into transformation, identified 60 overlapping included a two major events: lymphomagenesis with the t(11;14) translo- number of members of the p53 interaction network (CDC2, cation (initiation), and evolution into a more aggressive form BIRC5 and FOXM1), which is known to mediate (transformation). Unfortunately, it is still unknown which progression during the G2/M transition. Thus, we observe genes contribute to each event. In this study, we performed that the Wnt signaling pathway may play an important role in cDNA microarray experiments designed based on the concept initial lymphomagenesis in addition to t(11;14) translocations, that morphologically heterogeneous MCL samples would and that specific mitotic regulators facilitate transformation provide insights into the role of aberrant into more aggressive forms. for both events. A total of 15 MCLs were collected from the files, which include a total of 237 MCL patients confirmed Introduction by histology as CCND1-positive. We posited four stepwise morphological grades for MCL: MCL in situ, MCL with The pathological category of low-grade B cell non-Hodgkin's classical form (cMCL), MCL with aggressive form (aMCL), lymphoma (NHL) encompasses mature B cell neoplasms with and MCL with intermediate morphology between classical small to medium-sized cells such as follicular lymphoma (FL), and aggressive forms at the same site (iMCL). To identify marginal zone lymphoma (MZL), B cell chronic lymphocytic genes involved in initiation, we compared the tumor cells of leukemia/small lymphocytic lymphoma (B-CLL/SLL), and MCL in situ (n=4) with normal mantle zone B lymphocytes mantle cell lymphoma (MCL). The histological transformation (n=4), which were selected by laser microdissection (LMD). of low-grade lymphomas is thought to have a major impact To identify genes contributing to transformation, we selected on their prognosis, however, the clonal relationship between the overlapping genes differentially expressed between both the two neoplasm types and the pathogenesis underlying the cMCL (n=4) vs. aMCL (n=5) and classical vs. aggressive progression of the disease are still controversial. areas in iMCL (n=2) obtained by LMD. A significant number MCL is considered a well categorized B cell NHL because of genes (n=23, p=0.016) belonging to the Wnt signaling virtually all cases feature the chromosomal translocation pathway were differentially expressed in initiation. This t(11;14)(q13;q32) (IgH/CCND1) which leads to D1 specific activation was confirmed by immuno­histochemistry, (CCND1) overexpression. Despite this strong genetic hallmark, as MCL in situ had nuclear localization of phosphorylated- MCL presents in a broad spectrum of morphological forms with a high number of additional chromosomal abnormalities. From a practical point of view, we can detect the mantle zone growth pattern, which represents the initial component of MCL, in only 8.6% of cases (1). Aggressive transformation in Correspondence to: Dr Yoshizo Kimura, Department of Pathology, MCL is more notable, occurring in 32% of living cases and in School of Medicine, Kurume University, Asahi-machi 67, Kurume, Fukuoka 830-0011, Japan 70% of autopsy cases (2). Hence, the presence of pathological E-mail: [email protected] evidence such as characteristic cytologies (small, classical, blastic and pleomorphic) or proliferating patterns (mantle- Key words: mantle cell lymphoma, initiation, transformation, cDNA zone, nodular and diffuse) at the time of biopsy seems to microarray analysis, laser microdissection, Wnt signaling pathway, depend on temporal changes in the accumulation of molecular mitotic regulator genetic alterations. For an accurate understanding of MCL, there are some questions requiring resolution. First, why is this lymphoma 458 KIMURA et al: GENE EXPRESSION ANALYSIS OF MANTLE CELL LYMPHOMA

Table I. Data of studied patients.

Case Age Sex Diagnosis Tissue Growth pattern PI (Ki-67) IgH/CCND1 RNA extraction

R1 36 F Benign lymphadenitis LN - ND ND LMD R2 25 F Benign lymphadenitis LN - ND ND LMD R3 54 M Benign lymphadenitis LN - ND ND LMD R4 14 M Benign lymphadenitis LN - ND ND LMD MCL112 68 M MCL in situ LN MZ 15 + LMD MCL113 49 M MCL in situ LN MZ 10 + LMD MCL114 73 M MCL in situ LN MZ 20 + LMD MCL71 57 M MCL in situ LN MZ 5 + LMD MCL126 79 M Classical MCL (cMCL) LN N 25 + Whole MCL135 59 M cMCL LN N 20 + Whole MCL141 75 M cMCL LN N 25 + Whole MCL200 75 M cMCL LN N 30 + Whole MCL80 80 M Intermediate MCL (iMCL) LN N&D 20/80 + LMD MCL132 67 F iMCL LN N&D 15/95 + LMD MCL5 71 M Aggressive MCL (aMCL) LN D 70 + Whole MCL40 77 F aMCL LN D 70 + Whole MCL98 87 M aMCL LN D 95 + Whole MCL102 77 M aMCL LN D 75 + Whole MCL107 76 M aMCL LN D 60 + Whole

LMD, laser microdissection; PI, proliferation index; ND, not done; LN, lymph node; MZ, mantle zone; N, nodular; D, diffuse.

characterized by biological heterogeneity despite the charac- Materials and methods teristic translocation event? Second, what is required for the initial pathogenesis of MCL in addition to the t(11;14) translo- Patients and tissue samples. We performed cDNA microarray cation? A recent study has shown that the t(11;14) translocation experiments using frozen tissues of 19 lymph node biopsies. A alone is not sufficient to produce tumors (3) and, though at total of 15 MCLs were collected from the files of the Departments very low levels, t(11;14)-positive cells have been found in the of Pathology of Kurume University (Fukuoka, Japan), which blood of healthy individuals (4). Thus, key molecular behavior include a total of 237 MCL patients (1) confirmed by histology as in MCL is thought to be divided into two major stages: CCND1-positive (Table I). All 15 cases were subjected to cytoge- initial lymphomagenesis in addition to the t(11;14) transloca- netic and/or FISH studies and found positive for the IgH/CCND1. tion (initiation) and the accumulation of variable secondary This study was approved by the Kurume University Institutional genomic alterations occurring over time, leading to the evolu- Review Board, and patients provided informed consent in accor- tion into more aggressive forms (transformation). dance with the Declaration of Helsinki. Recently, microarray profiling studies have identified In order to identify differentially expressed genes and their differential expression of several genes in the progression of contribution to each event, we posited four stepwise morpho- MCL (5-7). Some of these studies have used purified mantle logical grades for MCL: MCL in situ, MCL with classical form zone B lymphocytes sorted from reactive tonsillectomy speci- (cMCL), MCL with aggressive form (aMCL), and MCL with mens for gene expression profiling to avoid contamination with intermediate morphology between classical and aggressive stromal cells, T cells and macrophages (6,7). In these studies, forms at the same site (iMCL) (Fig. 1). Namely, MCL in situ is several genes identified as overexpressed in aggressive form defined as those samples with a very thin neoplastic mantle zone are known to be involved in cell cycle control or apoptotic cell growth pattern and very little or no spreading of tumor cells death, and several genes related to the PI3K/AKT, WNT and into interfollicular areas. cMCL is characterized by a prominent TGF-β signaling pathways are reportedly important in the nodular proliferation of atypical medium-sized tumor cells pathogenesis of MCL. Although morphological heterogeneity (classical form) without diffuse proliferation. aMCL is desig- remains a central issue in the prognosis and treatment of nated as a combination of two morphological forms: blastoid this lymphoma, no gene expression profiling studies to date and pleomorphic. iMCL contains distinct areas of both classical have attempted to identify the genes and signaling pathways and aggressive forms. involved in the each event separately. The present study was For evaluation of initiation, we compared samples from the designed based on the concept that gene expression profiling of tumor cells of MCL in situ (n=4) with those from normal mantle morphologically heterogeneous MCL samples would provide zone B-lymphocytes from benign lymphadenitis (n=4). Samples insight into the role of aberrant gene expression for both initial were derived from selected specimens by means of laser lymphomagenesis and transformation events. microdissection (LMD). We hypothesize that transformation INTERNATIONAL JOURNAL OF ONCOLOGY 43: 457-468, 2013 459

Figure 1. Schema of the stepwise morphological grades for MCL used in the present study. This diagram shows the morphological evolution of MCL occurring in parallel with transformational development (classical, intermediate, and aggressive) and expansive development (mantle zone, nodular, nodular & diffuse, and dif- fuse). The origin of MCL cells is postulated as naive CD5-positive pre-germinal center B cells. In order to identify the genes contributing to each event, we posited four stepwise morphological grades for MCL: MCL in situ, MCL with classical form (cMCL), MCL with aggressive form (aMCL), and MCL with intermediate morphology between classical and aggressive forms at the same site (iMCL). MCL in situ represents the mantle zone growth pattern in that the neoplastic mantle zone was very thin and there was very little if any spreading of tumor cells into interfollicular areas. cMCL is characterized by a prominent nodular proliferation of atypical medium-sized tumor cells (classical form) without diffuse area. iMCL contains both classical and aggressive form areas but separated by border lines (black dotted line). aMCL was designated a combination of two morphological forms, blastoid and pleomorphic.

into aMCL is a multistep process, similar to the progression of stained rapidly with Toluidine Blue O (Chroma-Gesellschaft carcinomas. It may be beneficial to subdivide aMCL neoplasms Schmid GmbH & Co., Köngen, Germany) and then washed into two major subtypes: ‘aMCL with genetic stepwise process’ in DEPC-treated water and air-dried with a fan. The frozen which results from the accumulation of many molecular genetic sections were microdissected with a Leica LMD6000 laser alterations and ‘de novo aMCL’ resulting from fewer but stronger microdissection system by following the company's protocol genetic alterations. Therefore, comparing cMCL with aMCL is (Leica, Wetzlar, Germany). The sorting regions were micro- not sufficient for the detection of transformation specific genes dissected from the tissue sections with LMD (Fig. 3), and the because it is not possible to distinguish them morphologically. dissected cells were collected in 0.5 ml tubes filled with 50 µl Alternatively, we expected that using results from the gene lysis buffer for RNA extraction. selection method in combination with that of iMCL cases (n=2) would provide the candidate genes restricted within aMCL RNA extraction and biotinylated cRNA amplification. Total having genetic stepwise process. Additionally, using iMCL RNA was extracted from the LMD-obtained samples with samples as a discovery set can compensate for individual an RNAqueous-Micro kit (Ambion, Austin, TX, USA) differences between cases. For a concrete explanation of trans- according to the manufacturer's instructions. For cMCL and formation, total RNA from whole tumor tissue samples, cMCL aMCL tissues, total RNA was isolated using TRIzol reagent (n=4) and aMCL (n=5) were used [Fig. 2(I)]. We also compared (Invitrogen, Carlsbad, CA, USA). RNA samples were quantified classical areas with aggressive areas in iMCL obtained by LMD by an ND-1000 spectrophotometer (NanoDrop Technologies, [Fig. 2(II)]. Finally, we selected the overlapping genes differen- Wilmington, DE, USA) and the quality was confirmed with an tially expressed in both comparisons [Fig. 2(III)]. Experion System (Bio-Rad Laboratories, Hercules, CA, USA). Both cRNA amplification and labeling with biotin were used Laser microdissection (LMD). The tissue samples were to prepare samples for gene expression profiling by microarray immediately frozen in acetone/dry ice and stored at -80˚C for analysis. Briefly, 350-500 ng total RNA was amplified over- microdissection. The lymph node samples were embedded night (14 h) with the Illumina Total Prep RNA Amplification in an optical cutting temperature (OCT) compound (Sakura kit (Ambion) in accordance with the manufacturer's protocol. Finetek, Tokyo, Japan) and frozen in liquid nitrogen. Reaction cRNA was biotinylated during in vitro . Cryosections (10-µm-thick) were mounted on 2.0-µm-thick PEN-Membrane slides (MicroDissect GmbH, Herborn, Illumina BeadChips microarray. Sentrix Human WG-6 v3.0 Germany). After fixation in 100% ethanol, the slides were Expression BeadChips were purchased from Illumina, Inc. 460 KIMURA et al: GENE EXPRESSION ANALYSIS OF MANTLE CELL LYMPHOMA

Figure 2. Gene selection methods for the transformation. For the analysis of the method (I), we compared total RNAs from whole tumor tissue samples, cMCL (n=4) with aMCL (n=5). To compensate for individual differences and the influence ofde novo aMCL, we also compared microdissected classical area with the aggressive area in iMCL (n=2) obtained by means of LMD [method (II)]. Finally, we selected the overlapping genes (III) differentially expressed in both comparisons.

Figure 3. Microdissection of the present cases (3a and 3b). After Toluidine blue staining, a section was subjected to LMD. Figures showing microdissected areas of (a) classical form in iMCL (case no. MCL132) and (b) of tumor cells in MCL in situ (case no. MCL113), respectively.

(San Diego, CA, USA). More than 48,000 different bead types, Data analysis and filter criteria.Pre-processing was performed each with a 50-base gene-specific probe, are represented on on the raw signal intensities of all samples by log2-trans­ a single Beadchip. For each probe represented on the array, formation and normalization using the quantile algorithm of beads are assembled with an average 30-fold redundancy. A the ‘preprocess Core’ library package (8) in Bioconductor (9). hybridization mixture containing 1.5 µg biotinylated cRNA We selected the probes (excluding the control probes) where was hybridized to the beadchips at 58˚C overnight (18 h) the detection p-value was <0.01 in all samples, and used only before being washed and stained with streptavidin-Cy3 these probes to identify differentially expressed genes. For (GE Healthcare, Buckinghamshire, UK) according to the the samples meeting the criteria of initiation (MCL in situ vs. manufacturer's protocol. Beadchips were scanned on Illumina normal mantle zone B lymphocytes) and transformation (I) BeadStation 500 and fluorescent hybridization signals were (cMCL vs. aMCL), we applied Linear Models for Microarray assessed with Illumina BeadStudio software. Analysis (limma) package (10) of Bioconductor. We used a cutoff of limma p-value <0.05 and absolute log-fold change Immunohistochemistry. The samples were also evalu- (|logFC|) >0.5 to assess differentially expressed genes in each ated for expression of markers such as Wnt3 (Abcam comparison. For the transformation (II) comparison (classical Inc., Cambridge, MA, USA), phosphorylated-β-catenin area vs. aggressive area in iMCL), we selected probes that (Ser552) (pβ-catenin-S552) (Cell Signaling, Boston, MA, contained a ‘P’ flag in both iMCL samples. To identify up or USA), and Ki-67 (Dako Cytomation, Glostrup, Denmark). downregulated genes between aggressive and classical areas, we Formalin-fixed, paraffin-embedded tissues were used for all calculated Z-scores (11) and ratios (non-log scaled fold change) immunohistochemical stains. Antibody dilutions and antigen from the normalized signal intensities of each probe. Then we retrieval procedures were performed as standards. established criteria for regulated genes: absolute Z-score (|Z|) INTERNATIONAL JOURNAL OF ONCOLOGY 43: 457-468, 2013 461

Figure 4. Immunohistochemistry of (a, b, e and f) pβ-catenin-S552 and (c, d, g and h) Wnt3. The immunohistochemical studies revealed that the tumor cells of the MCL in situ showed nuclear localization of pβ-catenin-S552 (a and b) with high levels of cytoplasmic Wnt3 staining (c and d). The reactive mantle zone B cells were negative for pβ-catenin-S552 in the nuclear (e and f) with low lelels of cytoplasmic Wnt3 staining (g and h). pβ-catenin-S552 and Wnt3, original magnification: a, c, e and g, x1,000 and b, d, f and h, x2,000. GC, germinal center; MZ, marginal zone.

>2. Ingenuity Pathway Analysis (IPA6.0; Ingenuity Systems, experiments are measuring biologically relevant differences Redwood, CA, USA; http://www.ingenuity.com) was used to between the sample types. The top 10 up- and down-regulated identify networks of interacting genes. Lists of expressed (up genes in MCL in situ were: up, CCND1, FCGBP, IL17RB, and downregulated) genes were uploaded for IPA. A heat map NULL (probe ID; 5390192), WNT3, D4S234E, FBLN2, of the differentially expressed genes was generated by MeV CPXM1, DBN1 and TEAD2; down, PLAC8, LOC439949, software (12). Array data are available on the Gene Expression HVCN1, FGR, LOC651751, IL7R, DEF8, TXNDC8, GZMB Omnibus (GEO) website under accession numbers GSE30189 and SLAMF1. (http://www.ncbi.nlm.nih.gov/geo/). For further analysis, we focused on the canonical β-catenin- dependent Wnt pathway (WCP), which is critically involved Results in cell fate and diffrerentiation (13). As reasons for this, IPA analysis revealed that a significant number of differentially Differentially expressed genes in initiaion. In the analysis expressed genes (n=23) belong to this category (p=0.016), and of MCL initiation, we selected 1,538 genes (851 upregulated none of the genes in this pathway were observed as differen- and 787 downregulated) which showed significant differences tially expressed in the transformation experiments (Table II). (p<0.05) between normal mantle zone B lymphocytes and Whether the cytosolic pool of β-catenin participates in WCP MCL in situ samples (http://www.ncbi.nlm.nih.gov/geo/). The signaling is dictated by the availability of its binding partners, identification of CCND1 as the most significantly upregulated and these binding interactions are regulated by phosphoryla- gene in ‘initiation’ samples gives us confidence that these array tion (14). Phosphorylation of β-catenin at Ser552 by AKT can 462 KIMURA et al: GENE EXPRESSION ANALYSIS OF MANTLE CELL LYMPHOMA

Table II. Genes associated with Wnt signaling pathway (p<0.05).

Time course p<0.05 ------Normal MZ vs. MCL in situ cMCL vs. aMCL Acc Gene name Gene description Fold change P-value Fold change P-value

NM_053056.2 CCND1 Cyclin D1 4.41 2.51E-08 NS NS NM_030753.3 WNT3a Wingless-type MMTV 2.65 0.000063 NS NS integration site family, member 3 NM_024507.2 KREMEN2 Kringle containing transmembrane 1.52 0.001595 NS NS 2 NM_002335.1 LRP5b Low density lipoprotein 1.04 0.000428 NS NS receptor-related protein 5 NM_020335.1 VANGL2 Vang-like 2 (van gogh, Drosophila) 1.00 0.000208 NS NS NM_001466.2 FZD2 Frizzled homolog 2 (Drosophila) 0.90 0.001185 NS NS NM_003199.2 TCF4b Transcription factor 4 0.81 0.000155 NS NS NM_004380.2 CREBBPb CREB binding protein 0.72 0.000165 NS NS NM_001760.2 GRN Granulin 0.67 0.014332 NS NS NM_138713.2 NFAT5 Nuclear factor of activated T-cells 5, 0.57 0.001586 NS NS tonicity-responsive NM_002738.5 PRKCB Protein kinase C, beta 0.55 0.003660 NS NS NM_152221.2 CSNK1Eb Casein kinase 1, epsilon 0.53 0.004865 NS NS NM_001006610.1 SIAH1 Seven in absentia homolog 1 0.52 0.001216 NS NS (Drosophila) NM_004423.3 DVL3b Dishevelled, dsh homolog 3 0.45 0.007706 NS NS (Drosophila) NM_020248.2 CTNNBIP1 Catenin, beta interacting protein 1 0.42 0.016759 NS NS NM_003200.1 TCF3 Transcription factor 3 (E2A immunoglobulin 0.35 0.031493 NS NS enhancer binding factors E12/E47) NM_181492.1 TCF20 Transcription factor 20 (AR1) -0.39 0.020984 NS NS NM_033120.2 NKD2 Naked cuticle homolog 2 (Drosophila) -0.39 0.037969 NS NS NM_002872.3 RAC2 Ras-related C3 botulinum toxin substrate 2 -0.42 0.008708 NS NS (rho family, small GTP binding protein Rac2) NM_007236.3 CHP Calcium binding protein P22 -0.45 0.005099 NS NS NM_002467.3 MYC V-myc myelocytomatosis viral oncogene -0.48 0.034874 NS NS homolog (avian) NM_002737.2 PRKCA Protein kinase C, alpha -0.72 0.002025 NS NS NM_016269.2 LEF1 Lymphoid enhancer-binding factor 1 -0.87 0.004964 NS NS aMCL signature genes in agreement with CCND1-negative lymphomas that had MCL morphology and immunophenotype were classified as MCL (17). bGenes aberrantly expressed in MCL related to WNT signaling pathway compared with naive B cells (6).

enhance β-catenin/TCF reporter activation, suggesting that and aMCL groups [Fig. 2(I)] (http://www.ncbi.nlm.nih.gov/ pβ-catenin-S552 is a nuclear-localized form of β-catenin (active geo/). Gene expression profiling in iMCL showed that 220 genes form) (15). Microarray results indicate that β-catenin was not were differently expressed between microdissected classical significantly upregulated in MCL initiation, but immuno­ area and aggressive area samples [Fig. 2(II)] (http://www.ncbi. histochemical results revealed that the tumor cells of the MCL nlm.nih.gov/geo/). Finally, we identified 60 overlapping genes in situ samples showed nuclear localization of pβ-catenin-S552 [Fig. 2(III)] (Table III), both upregulated (n=58) and down- with high levels of cytoplasmic Wnt3 staining (Fig. 4a-d). On regulated (n=2), which showed significant differences in pre- vs. the other hand, reactive mantle zone B cells were negative for post-transformation with a p-value of <0.05. These genes are pβ-catenin-S552 in the nuclear, and cytoplasmic Wnt3 staining visualized as a heat map in Fig. 5. was weak (Fig. 4e-h). As may be expected, IPA analysis of the filtered 60 genes revealed that most of these genes (42/60; 70%) were classified Differentially expressed genes in transformation. With the in the following categories, all of which are thought to be Illumina BeadStudio software, we detected 710 genes that involved in transformation: 36 genes belonged to the category showed significant differences in expression between cMCL of ‘Cell cycle progression’ (UBE2C, BIRC5, CDCA5, TYMS, INTERNATIONAL JOURNAL OF ONCOLOGY 43: 457-468, 2013 463

Figure 5. Heat map of the differentially expressed genes in the transformation. The genes displayed in this figure were identified by our selection methods. Color intensity represents the magnitude of deviation from the median. Yellow represents relative expression levels higher than the median level across all tissues and blue represents lower expression levels. The levels of most genes are more strongly upregulated in aggressive compared to classical MCL samples.

KIFC1, TOP2A, FOXM1, CCNF, , KIF2C, TPX2, KIF11, highly expressed genes in aMCL, which suggests that they may CDC2, CENPF, NEK2, CDKN3, PTTG1, DLGAP5, CENPE, play a critical role in transformational events along with p53. CDC25A, AURKA, CCNB2, ASPM, KIF20A, UHRF1, Of these transformation-associated genes, only CDCA5 and CEP55, NCAPG, MKI67, KIF23, H2AFX, GPSM2); 28 MKI67 were also significantly changed in initiation, while the ‘DNA replication, recombination and repair’ genes (AURKA, remaining 58 genes were aberrantly expressed specifically in BIRC5, CENPE, DLGAP5, KIF2C, KIFC1, NCAPG, TOP2A, transformation. CCNB2, GPSM2, NEK2, PTTG1, KIF11, TPX2, CDCA5, CDK1, CDC25A, E2F2, FEN1, FOXM1, HMGB2, KIAA0101, Discussion H2AFX, POLQ, TYMS, KIF23, CENPA, UHRF1); 10 ‘cell death’ genes (A4GALT, AURKA, BIRC5, CCL2, CDC25A, Biological heterogeneity has become an important position CDCA2, CDK1, E2F2, NEK2, PTTG1, TOP2A, TYMS, in the current understanding of MCL. To meet this need, we UBE2C, CENPF, FEN1, FOXM1, HMGB2, KIF2C). As seen performed cDNA microarray experiments stratifying MCL in Fig. 6, these genes recapitulate much of the p53 interaction samples into four morphological grades based on the files network, a network centrally involved in cell cycle progres- including a large series of patients. In order to investigate sion. Most of these genes (AURKA, BIRC5, CDC2, CDC25A, the molecular mechanisms underlying these differences, we CENPF, CENPE, CCNB2, FOXM1, NEK2, PTTG1, TPX2, compared initial component samples or transformational ASPM, TOP2A, DLGAP5, KIF2C, KIF23, UBE2C) were morphology samples and identified genes associated with the considered important players in MCL transformation because Wnt signaling pathway and several known mitotic regulators as of their known function as mitotic regulators, which mediate differentially regulated. cell cycle progression during the G2/M transition. CDC2, Cell cycle alterations resulting in unscheduled proliferation FOXM1 and BIRC5 in particular interact with many other are strongly associated with the evolution of malignant tumors. 464 KIMURA et al: GENE EXPRESSION ANALYSIS OF MANTLE CELL LYMPHOMA

Table III. Genes associated with the development of aggressive form of MCL (cMCL vs. aMCL using by iMCL).

Acc Gene name Gene description Fold change P-value

NM_181800.1 UBE2C -conjugating enzyme E2C 1.84 0.000371 NM_181803.1 UBE2C Ubiquitin-conjugating enzyme E2C 1.74 0.000558 NM_002982.3 CCL2 Chemokine (C-C motif) ligand 2 1.60 0.038047 NM_001168.2 BIRC5d,e Baculoviral IAP repeat-containing 5 1.58 0.000489 NM_080668.2 CDCA5 cycle associated 5 1.27 0.003079 NM_004701.2 CCNB2c Cyclin B2 1.21 0.004265 NM_001071.1 TYMSc Thymidylate synthetase 1.11 0.001068 NM_002263.2 KIFC1 family member C1 1.09 0.007440 NM_001067.2 TOP2Aa,c Topoisomerase (DNA) II alpha 170 kDa 1.08 0.006076 NR_002734.1 PTTG3 Pituitary tumor-transforming 3 1.07 0.014845 NM_018136.3 ASPMb,c,d,e ‘Abnormal spindle (ASP) homolog, 1.07 0.005186 microcephaly associated (Drosophila)’ NM_021953.2 FOXM1c Forkhead box M1 1.01 0.003420 NM_005733.1 KIF20Ac Kinesin family member 20A 0.99 0.010183 NM_001761.1 CCNF Cyclin F 0.99 0.006634 NM_001048201.1 UHRF1a,e Ubiquitin-like with PHD and ring finger domains 1 0.97 0.007291 NM_001002876.1 CENPM protein M 0.94 0.016725 NM_004091.2 E2F2 transcription factor 2 0.94 0.019567 NM_001002876.1 CENPM Centromere protein M 0.94 0.035235 NM_030919.2 FAM83D ‘Family with sequence similarity 83, member D’ 0.92 0.007561 NM_024053.3 CENPM Centromere protein M 0.92 0.046319 NM_006845.2 KIF2C Kinesin family member 2C 0.90 0.016135 NM_199420.3 POLQ ‘Polymerase (DNA directed), theta’ 0.89 0.004755 NM_012112.4 TPX2 ‘TPX2, -associated, 0.89 0.016590 homolog (Xenopus laevis)’ NM_018131.3 CEP55 Centrosomal protein 55 kDa 0.88 0.006728 NM_004523.2 KIF11c Kinesin family member 11 0.88 0.005905 NM_022346.3 NCAPG ‘Non-SMC condensin I complex, subunit G’ 0.88 0.014421 XM_934471.1 LOC399942 - 0.87 0.039823 NM_002417.3 MKI67e Antigen identified by monoclonal antibody Ki-67 0.86 0.022646 NM_001786.2 CDC2b,c,d,e ‘Cell division cycle 2, G1 to S and G2 to M’ 0.86 0.023818 NM_016343.3 CENPFb,c,e ‘Centromere protein F, 350/400ka (mitosin)’ 0.84 0.032315 NM_002497.2 NEK2c NIMA (never in mitosis gene a)-related kinase 2 0.83 0.002697 NM_178448.2 C9orf140 9 open reading frame 140 0.81 0.023110 NM_004219.2 PTTG1 Pituitary tumor-transforming 1 0.80 0.047544 NM_004856.4 KIF23c,e Kinesin family member 23 0.80 0.007960 NM_005192.2 CDKN3a,c Cyclin-dependent kinase inhibitor 3 0.79 0.029828 NM_004219.2 PTTG1 Pituitary tumor-transforming 1 0.79 0.038189 NM_002129.2 HMGB2a,c High-mobility group box 2 0.79 0.045729 NM_152515.2 CKAP2L associated protein 2-like 0.78 0.036949 NM_004523.2 KIF11c Kinesin family member 11 0.78 0.004380 NM_014746.2 RNF144A Ring finger protein 144A 0.78 0.005460 NM_014750.3 DLGAP5 ‘Discs, large (Drosophila) 0.76 0.009518 homolog-associated protein 5’ NM_014736.4 KIAA0101c KIAA0101 0.76 0.045005 NM_032704.2 TUBA1Cb ‘, alpha 1c’ 0.73 0.027816 NM_001813.2 CENPEc,d ‘Centromere protein E, 312 kDa’ 0.73 0.013906 BC004287 NULL Null 0.73 0.019170 NM_001789.2 CDC25A Cell division cycle 25 homolog A (S. pombe) 0.73 0.025460 NM_002105.2 H2AFX ‘H2A histone family, member X’ 0.72 0.038540 NM_203394.2 E2F7 E2F transcription factor 7 0.68 0.032400 NM_001042426.1 CENPAc Centromere protein A 0.65 0.044459 NM_198434.1 AURKAc Aurora kinase A 0.62 0.042391 INTERNATIONAL JOURNAL OF ONCOLOGY 43: 457-468, 2013 465

Table III. Continued.

Acc Gene name Gene description Fold change P-value

NM_152562.2 CDCA2 Cell division cycle associated 2 0.61 0.040259 NM_006082.2 TUBA1Ba,f ‘Tubulin, alpha 1b’ 0.61 0.002775 NM_004111.4 FEN1 Flap structure-specific endonuclease 1 0.60 0.035752 NM_013296.3 GPSM2 ‘G-protein signaling modulator 2 0.55 0.033270 (AGS3-like, C. elegans)’ NM_182513.1 SPC24 ‘SPC24, NDC80 kinetochore complex component, 0.55 0.034394 homolog (S. cerevisiae)’ XR_016048.1 MGC40489 - 0.53 0.049409 NM_001733.4 C1R ‘Complement component 1, R subcomponent’ 0.52 0.039100 NM_006461.3 SPAG5c Sperm associated antigen 5 0.48 0.008955 NM_173509.2 FAM163A ‘Family with sequence similarity 163, member A’ -0.39 0.007364 NM_017436.4 A4GALT ‘Alpha 1,4-galactosyltransferase’ -1.28 0.000604 aHigh level of proliferation signature genes in MCL (17). bProliferation signature genes that predicted length of survival with high statistical significance in MCL (17). cIdentification of genes diffferentally expressed in MCL with low levels of long CCND1 transcript compared to other MCL (fold up) (18). dSurvival prediction genes in MCL (20). eMost significant genes separating good and poor prognosis in MCL (fold up) (20). fGenes separating good and poor prognosis not associated with cell cycle (fold up) (20).

Figure 6. Gene network generated through the use of IPA. IPA was applied to 60 genes obtained by using iMCL as the transformation to create a network interre- lated with mediator genes. Interacting nodes are defined by either direct relationships, which require direct physical contact (solid arrows), or indirect relationships (dashed arrows). The direction of the arrows shows the direction of the interaction. The chart clearly and in detail showed the network composed of p53 and several of its interaction partners. Most of these are mitotic regulators, and CDC2, FOXM1 and BIRC5 in particular interact with many other highly expressed genes in aMCL. The blue and orange lines indicate the connection with CDK1 (CDC2) and TP53, respectively. Red molecules indicate higher expression in aMCL than in cMCL at gene level. 466 KIMURA et al: GENE EXPRESSION ANALYSIS OF MANTLE CELL LYMPHOMA

Most of these changes alter pathways involved in G1 progres- transcription factor, and is intimately involved in mitosis sion or the G1/S transition (16). With regard to MCL, the regulation (25). Namely, FOXM1 directly or indirectly (via ‘Lymphoma and Leukemia Molecular Profiling Project MYC) regulates genes that control G1/S transition, S-phase (LLMPP)’ demonstrated that the length of survival of MCL progression, G2/M transition and M-phase progression (25), patients depends upon quantitative differences in progression supporting that it is involved in tumorigenesis. It is of interest from G1/S phase of the cell cycle (17). Sander et al also showed that its expression level is known to be increased in several that a subset of MCL tumors with low levels of the long CCND1 other tumor grades, such as prostate carcinoma and glioblas- transcript is highly proliferative and some of its related genes toma (26,27). Thus, all of these genes, as well as the other have homology to the group of cell cycle (G1/S) promoting E2F mitotic regulators we found to have differential expression transcription partners (18). Indeed, there is a strong correlation in MCL transformation, are potentially attractive therapeutic between these reported genes and the 60 we identified in this targets or strong diagnostic tools. study as involved in transformation (Table III), which empha- Alterations in the DNA damage response pathway and sizes the importance of G1/S regulators in MCL evolution. The mitotic checkpoints through p53 are major additional genetic morphological classifications used in this study were based on events in MCL, as indicated by the high rate of tetraploidy our previous findings that MCL presents in three morphological found in aMCL (28). In fact, p53 inactivation in MCL as a evolutions: classical, intermediate and aggressive, and that the consequence of deletion or mutation occurs more frequently aggressive form is considered an important clinical signal of in the aggressive form than in the classic form (29). However, poor prognosis (1). Therefore, it might be natural to expect that inactivating mutations of p53 are found in only 38% of dysregulation of G1/S proliferation signature genes is neces- aMCL cases (30,31), suggesting that several other genetic sary to determine both the length of MCL patient survival and alterations also contribute to inactivation of the p53 pathway. the time to neoplasm transformation. Currently, ATM deletions are thought to be strongly associ- Although the dysregulation of progression through mitosis ated with the dysregulation of the DNA damage response does not directly promote proliferation, a few centrosomal pathway through p53, and are probably present in the early and mitotic (such as AURKA, PLK1 and PTTG1) phase of MCL (17,32‑34). Our microarray experiments have have been reported to act as oncogenes (19). Interestingly, our revealed that ATM was significantly downregulated in initia- results highlight a large number of mitotic regulators (CDC2, tion with strong fold change (FC=-0.73), while there was no AURKA, BIRC5, CDC25A, CENPE, CCNB2, FOXM1, NEK2, significant change observed in transformation. In the results PTTG1, TPX2, DLGAP5, KIF2C, KIF23 and UBE2C) in of our previous study using immunohistochemistry, higher addition to G1/S regulators (CDC25A, E2F2, FOXM1 and cell positivity of p53 (DO7) was observed in iMCL than in TYMS). Of these genes, CDC2, FOXM1 and BIRC5 interact cMCL, while there was no significant difference between with many other highly expressed mitotic regulators within iMCL and aMCL (1). This is due to the fact that there is a the p53 interaction network, suggesting that specific mitotic high concordance between p53 (DO7) nuclear overexpression regulators facilitate the transformation of MCL into more and gene mutation in human carcinomas (35). These results aggressive form. In support of this hypothesis, Blenk et al strongly indicated that dysregulation of the ATM/p53 pathway identified significantly differentially expressed genes between in MCL would occur at undetectable levels as a relatively samples with good and poor prognosis in MCL using explor- early phenomenon. After that, the disruption of central atory analysis of gene expression values and CGH data (20). mitotic regulators (CDC2, BIRC5 and FOXM1) is responsible Surprisingly, the majority of the genes highlighted in that for the induction of chromosomal instability. Combined, these study (BIRC5, ASPM, MKI67, UHRF1, CDC2, CENPF and factors play an important pathogenic role in the evolution of KIF23), including their survival predictor of genes for MCL MCL, perturbing the regulation of tumor cell cycling at the patients (CENPE, CDC2, BIRC5 and ASPM), overlap with G2/M transition. our current findings (Table III). These observations imply that The mechanisms of the initial lymphomagenesis of both CDC2 and BIRC5 expression play central roles in MCL MCL in addition to the IgH/CCND1 remain unclear, and transformation and thus reflect prognosis. new approaches are urgently needed to elucidate which CDC2 (also known as CDK1) is one of the master regula- genes and signaling pathways contribute to this event. Our tors of mitosis, as it is involved in the cycle and results support the hypothesis that aberrant Wnt3 signaling is early mitotic events. Hui et al reported that elevated protein required for the MCL lymphomagenesis, because a significant levels of CDC2 are correlated with the expression of prolif- number of WCP associated genes were aberrantly expressed eration marker, and represent a useful and simple method in in the initiation and not significantly changed in the transfor- evaluating the prognosis of MCL patients (21). Although BIRC5 mation. Moreover, immunohistochemical findings revealed (also known as survivin) is known as a member of inhibitor of the special activation of WCP. apoptosis proteins gene family (22), several studies have shown The significance of WCP signaling in tumor initiation may that its role in cancer is not limited to apoptosis inhibition (23). be straightforward from the view point of the adenoma-carci- Accordingly, BIRC5 expression levels are higher in aMCL noma sequence (36). Deletion of the APC gene resulting in the samples than cMCL samples and are associated with prolifera- activation of WCP is a consistent finding among the earliest tive activity and survival of the patients (24). events in both de novo and sporadic colon carcinomas. Also, The role of FOXM1 in lymphoma has not been reported in several studies support the notion that WCP can influence both the literature. However, FOXM1 was also significantly upreg- lymphopoiesis (37,38) and hematological malignancies (39). ulated genes in MCL transformation (FC; 1.01 and p<0.005) In our initiation experiments, especially Wnt3 and LRP5 in our results. FOXM1 is a typical proliferation-associated were genes with a high average increase (2.65- and 1.04-fold, INTERNATIONAL JOURNAL OF ONCOLOGY 43: 457-468, 2013 467 respectively), and are reportedly highly expressed in MCL 5. De Vos S, Krug U, Hofmann WK, et al: Cell cycle alterations in the blastoid variant of mantle cell lymphoma (MCL-BV) as cases (40). detected by gene expression profiling of mantle cell lymphoma The binding of Wnt proteins to their respective cell (MCL) and MCL-BV. Diagn Mol Pathol 12: 35-43, 2003. surface receptors, including seven of the transmembrane 6. Rizzatti EG, Falcao RP, Panepucci RA, et al: Gene expres- sion profiling of mantle cell lymphoma cells reveals aberrant fizzled (Fz) receptors and low-density lipoprotein receptor- expression of genes from the PI3K-AKT, WNT and TGFbeta related protein (LRP5 or 6), activates disheveled (Dvl). signalling pathways. Br J Haematol 130: 516-526, 2005. Activated Dvl can inhibit the degradation of β-catenin by 7. Martinez N, Camacho FI, Algara P, et al: The molecular signature of mantle cell lymphoma reveals multiple signals the destruction complex, which is composed of adenoma- favoring cell survival. Cancer Res 63: 8226-8232, 2003. tous polyposis coli (APC), axin, casein kinase 1 (CK1) and 8. Bolstad BM, Irizarry RA, Astrand M and Speed TP: A compar- GSK-3β. Consequently, accumulation of β-catenin in the ison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19: nucleus regulates gene expression in cooperation with T cell 185-193, 2003. factor (TCF)/lymphocyte enhancer factor (LEF) transcrip- 9. Gentleman RC, Carey VJ, Bates DM, et al: Bioconductor: open tion factors, resulting in the activation of the WCP target software development for computational biology and bioinfor- matics. Genome Biol 5: R80, 2004. genes such as CCND1 and c-Myc (13). 10. Smyth GK, Michaud J and Scott HS: Use of within-array Using gene expression profiling, Rosenwald et al identi- replicate spots for assessing differential expression in micro- fied MCL signature genes in CCND1-negative lymphoma array experiments. Bioinformatics 21: 2067-2075, 2005. 11. Quackenbush J: Microarray data normalization and transfor- cases classified as MCL by both morphology and IHC. Wnt3 mation. Nat Genet 32 (Suppl): 496-501, 2002. is one of these MCL signature genes, suggesting that Wnt3 12. Saeed AI, Sharov V, White J, et al: TM4: a free, open-source is more intrinsic to MCL than CCND1 (17). According to system for microarray data management and analysis. Biotechniques 34: 374-378, 2003. Gelebart et al, Wnt 3 is highly and consistently expressed in 13. Aguilera O, Munoz A, Esteller M and Fraga MF: Epigenetic MCL as detected by WCP-specific oligonucleotide arrays (40). alterations of the Wnt/beta-catenin pathway in human disease. Lako et al showed that Wnt3 protein can enhance haematopoi- Endocr Metab Immune Disord Drug Targets 7: 13-21, 2007. 14. Daugherty RL and Gottardi CJ: Phospho-regulation of etic commitment during in vitro differentiation of embryonic beta‑catenin adhesion and signaling functions. Physiology stem cells (41). From these studies, we speculate that aber- (Bethesda) 22: 303-309, 2007. rant expression of Wnt3 can emerge as potential activator of 15. Fang D, Hawke D, Zheng Y, et al: Phosphorylation of beta‑catenin by AKT promotes beta-catenin transcriptional lymphomagenesis in MCL because CCND1 itself is an onco- activity. J Biol Chem 282: 11221-11229, 2007. genic target of activated WCP. In addition, several studies have 16. Massague J: G1 cell-cycle control and cancer. Nature 432: shown that Wnt3 is highly expressed in B cell CLL (42,43) as 298-306, 2004. 17. Rosenwald A, Wright G, Wiestner A, et al: The proliferation gene well as in MCL. expression signature is a quantitative integrator of oncogenic To summarize, this study shed light on the mechanisms events that predicts survival in mantle cell lymphoma. Cancer of initiation and evolution in MCL. The resulting patterns Cell 3: 185-197, 2003. 18. Sander B, Flygare J, Porwit-Macdonald A, et al: Mantle cell of gene dysregulation in these evens strongly indicate that lymphomas with low levels of cyclin D1 long mRNA transcripts the Wnt signaling pathway plays a critical role in initial are highly proliferative and can be discriminated by elevated lymphomagenesis, and that specific mitotic regulators facili- cyclin A2 and cyclin B1. Int J Cancer 117: 418-430, 2005. 19. Perez de Castro I, de Carcer G and Malumbres M: A census of tate transformation into more aggressive forms. Our unique mitotic cancer genes: new insights into tumor cell biology and approach may contribute to future understanding of various cancer therapy. Carcinogenesis 28: 899-912, 2007. mature B cell lymphomas. These data hint at a novel system 20. Blenk S, Engelmann JC, Pinkert S, et al: Explorative data analysis of MCL reveals gene expression networks implicated for the classification of low-grade B cell neoplasms using the in survival and prognosis supported by explorative CGH expression levels of WCP genes and specific mitotic regulator analysis. BMC Cancer 8: 106, 2008. genes as markers for disease stage and predicted outcomes. 21. Hui D, Reiman T, Hanson J, et al: Immunohistochemical detection of cdc2 is useful in predicting survival in patients with mantle cell lymphoma. Mod Pathol 18: 1223-1231, 2005. Acknowledgements 22. Schimmer AD: Inhibitor of apoptosis proteins: translating basic knowledge into clinical practice. Cancer Res 64: 7183-7190, 2004. The authors would like to thank Konomi Takasu, Mayumi Miura 23. Mita AC, Mita MM, Nawrocki ST and Giles FJ: Survivin: key and Kanoko Miyazaki for their technical support. regulator of mitosis and apoptosis and novel target for cancer therapeutics. Clin Cancer Res 14: 5000-5005, 2008. 24. Martinez A, Bellosillo B, Bosch F, et al: Nuclear survivin References expression in mantle cell lymphoma is associated with cell proliferation and survival. Am J Pathol 164: 501-510, 2004. 1. Kimura Y, Sato K, Arakawa F, et al: Mantle cell lymphoma 25. Wierstra I and Alves J: FOXM1, a typical proliferation-associ- shows three morphological evolutions of classical, intermediate, ated transcription factor. Biol Chem 388: 1257-1274, 2007. and aggressive forms, which occur in parallel with increased 26. Liu M, Dai B, Kang SH, et al: FoxM1B is overexpressed in labeling index of cyclin D1 and Ki-67. Cancer Sci 101: 806-814, human glioblastomas and critically regulates the tumorige- 2010. nicity of glioma cells. Cancer Res 66: 3593-3602, 2006. 2. Kaleem Z, Wakoff AR, Smith RP and Hess JL: Blastic trans- 27. Kalin TV, Wang IC, Ackerson TJ, et al: Increased levels of formation of mantle cell lymphoma. Arch Pathol Lab Med 120: the FoxM1 transcription factor accelerate development and 577-580, 1996. progression of prostate carcinomas in both TRAMP and LADY 3. Bodrug SE, Warner BJ, Bath ML, Lindeman GJ, Harris AW transgenic mice. Cancer Res 66: 1712-1720, 2006. and Adams JM: Cyclin D1 transgene impedes lymphocyte 28. Ott G, Kalla J, Ott MM, et al: Blastoid variants of mantle cell maturation and collaborates in lymphomagenesis with the myc lymphoma: frequent bcl-1 rearrangements at the major translo- gene. EMBO J 13: 2124-2130, 1994. cation cluster region and tetraploid chromosome clones. Blood 4. Hirt C, Schuler F, Dolken L, Schmidt CA and Dolken G: Low 89: 1421-1429, 1997. prevalence of circulating t(11;14)(q13;q32)-positive cells in the 29. Pinyol M, Hernandez L, Cazorla M, et al: Deletions and loss peripheral blood of healthy individuals as detected by real-time of expression of p16INK4a and p21Waf1 genes are associated quantitative PCR. Blood 104: 904-905, 2004. with aggressive variants of mantle cell lymphomas. Blood 89: 272-280, 1997. 468 KIMURA et al: GENE EXPRESSION ANALYSIS OF MANTLE CELL LYMPHOMA

30. Hernandez L, Fest T, Cazorla M, et al: p53 gene mutations and 37. Reya T, O'Riordan M, Okamura R, et al: Wnt signaling regulates protein overexpression are associated with aggressive variants B lymphocyte proliferation through a LEF-1 dependent of mantle cell lymphomas. Blood 87: 3351-3359, 1996. mechanism. Immunity 13: 15-24, 2000. 31. Greiner TC, Moynihan MJ, Chan WC, et al: p53 mutations in 38. Staal FJ and Sen JM: The canonical Wnt signaling pathway mantle cell lymphoma are associated with variant cytology and plays an important role in lymphopoiesis and hematopoiesis. predict a poor prognosis. Blood 87: 4302-4310, 1996. Eur J Immunol 38: 1788-1794, 2008. 32. Bentz M, Plesch A, Bullinger L, et al: t(11;14)-positive mantle 39. Ge X and Wang X: Role of Wnt canonical pathway in hemato- cell lymphomas exhibit complex karyotypes and share simi- logical malignancies. J Hematol Oncol 3: 33, 2010. larities with B-cell chronic lymphocytic leukemia. Genes 40. Gelebart P, Anand M, Armanious H, et al: Constitutive activa- Cancer 27: 285-294, 2000. tion of the Wnt canonical pathway in mantle cell lymphoma. 33. Monni O, Oinonen R, Elonen E, et al: Gain of 3q and deletion of Blood 112: 5171-5179, 2008. 11q22 are frequent aberrations in mantle cell lymphoma. Genes 41. Lako M, Lindsay S, Lincoln J, Cairns PM, Armstrong L and Chromosomes Cancer 21: 298-307, 1998. Hole N: Characterisation of Wnt gene expression during the 34. Stilgenbauer S, Winkler D, Ott G, et al: Molecular charac- differentiation of murine embryonic stem cells in vitro: role of terization of 11q deletions points to a pathogenic role of the Wnt3 in enhancing haematopoietic differentiation. Mech Dev ATM gene in mantle cell lymphoma. Blood 94: 3262-3264, 103: 49-59, 2001. 1999. 42. Lu D, Zhao Y, Tawatao R, et al: Activation of the Wnt signaling 35. Soong R, Robbins PD, Dix BR, et al: Concordance between p53 pathway in chronic lymphocytic leukemia. Proc Natl Acad Sci protein overexpression and gene mutation in a large series of USA 101: 3118-3123, 2004. common human carcinomas. Hum Pathol 27: 1050-1055, 1996. 43. Rosenwald A, Alizadeh AA, Widhopf G, et al: Relation of gene 36. Vogelstein B, Fearon ER, Hamilton SR, et al: Genetic altera- expression phenotype to immunoglobulin mutation genotype tions during colorectal-tumor development. N Engl J Med 319: in B cell chronic lymphocytic leukemia. J Exp Med 194: 525-532, 1988. 1639‑1647, 2001.